Now that inherit_tags has been removed, 3.3 has the same functionality as 3.2: tags
are *never* cascaded from parent to child (but you tags do inherit from defaults
to a service and instanceof to a service).
This PR was merged into the 3.3-dev branch.
Discussion
----------
[Security] json login listener: ensure a json response is sent on bad request
| Q | A
| ------------- | ---
| Branch? | master (3.3)
| Bug fix? | yesish
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | N/A
| License | MIT
| Doc PR | N/A
I would have simply recommended to set the proper format when declaring the route:
```yml
# routing.yml
api_login:
path: /login
defaults: { _format: json }
```
but, since https://github.com/symfony/symfony/pull/22477 has been merged, and considering https://github.com/symfony/symfony/pull/22477#issuecomment-295897629:
> my point above regarding checking the content type is so that one could use form_login and json_login in parallel on the same routes and within the same firewall
we may consider setting the request format to json when throwing the `BadRequestHttpException`, so used conjointly with the TwigBundle, the exception is rendered using the `exception.json.twig` template.
ping @lsmith77
(An alternative would be to check the Accept header to set the request format to json if it's the preferred one instead of doing it each time we throw the exception. But Symfony never used such content negotiation AFAIK, and I think it's safe enough to assume someone sending json is expecting json as ouput for exceptions.)
Commits
-------
4427cf9157 [Security] json login listener: ensure a json response is sent on bad request
This PR was merged into the 3.3-dev branch.
Discussion
----------
[DI] Throw useful exception on bad XML argument tags
| Q | A
| ------------- | ---
| Branch? | 3.3
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #22525
| License | MIT
| Doc PR | -
I still think that the feature request in #22525 would make things better.
But at least, let's make thing fail loudly, instead of silently today, with the associated usual wtfs :)
Commits
-------
91828ecd17 [DI] Throw useful exception on bad XML argument tags
This PR was merged into the 3.3-dev branch.
Discussion
----------
[Security] add Request type json check in json_login
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no, unreleased feature
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR | -
follow up to https://github.com/symfony/symfony/pull/22425 to limit the `UsernamePasswordJsonAuthenticationListener` to only requests with appropriate JSON content type.
I am not entirely happy with this implementation but mostly because Symfony out of the box only provides very limited content type negotiation. I guess anyone that wants to tweak the content negotiation will simply need to ensure the Request::$format is set accordingly before the code is triggered.
Commits
-------
045a36b303 add Request type json check in json_login
This PR was merged into the 3.3-dev branch.
Discussion
----------
[Serializer] Add missing normalizer options constants
| Q | A
| ------------- | ---
| Branch? | master (3.3)
| Bug fix? | not really
| New feature? | yesish, but for 3.3 as those options were added on this branch and not released yet
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | https://github.com/symfony/symfony/pull/22537#discussion_r113719848
| License | MIT
| Doc PR | N/A
As seen in https://github.com/symfony/symfony/pull/22537#discussion_r113719848.
@dunglas : I'm not sure about the exposing the `key_type` option as a constant in `ArrayDenormalizer`/`AbstractObjectNormalizer`, as it looks more or less like a detail of the AbstractObjectNormalizer implementation, but anyway it should be in 3.2 if we add it, so I haven't included it here.
However, I wonder if this option shouldn't directly accept a string too, rather than just a `Symfony\Component\PropertyInfo\Type` instance if we want to consider this option "public"?
Commits
-------
b0c414f2c8 [Serializer] Add missing normalizer options constants
This PR was merged into the 3.3-dev branch.
Discussion
----------
[Serializer] Allow to pass csv encoder options in context
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | N/A
| License | MIT
| Doc PR | N/A
CSV contents typically are provided by one or many third-parties, not always allowing you to get control over the provided format. In case you need to import csv files with different formats, either you have to instantiate a decoder yourself/inject it instead of the main serializer instance, either you need another serializer instance with a differently configured csv encoder registered within.
This PR allows to configure any encoder option through the context, so you can keep injecting and using the same serializer instance.
Commits
-------
10a76aac15 [Serializer] Allow to pass csv encoder options in context
This PR was merged into the 3.3-dev branch.
Discussion
----------
[Security] Handle bad request format in json auth listener
| Q | A
| ------------- | ---
| Branch? | master (3.3)
| Bug fix? | yesish
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | N/A
| License | MIT
| Doc PR | N/A
In https://github.com/symfony/symfony/pull/22034, I wondered myself if we shouldn't throw a dedicated exception to handle bad formatted requests and give more inputs to the client by returning a 400 response with an explicit message.
~~Here is a suggestion, introducing a new `BadRequestFormatException` and using it in `UsernamePasswordJsonAuthenticationListener` whenever there is no custom failure handler set (but someone using its own handler should be able to treat the failure properly too).~~
As discussed with @chalasr , it seems better to directly throw a `BadRequestHttpException` as it's actually out of the whole security process. PR updated.
Commits
-------
93a8cb9cd4 [Security] Handle bad request format in json auth listener
This PR was squashed before being merged into the 3.3-dev branch (closes#22551).
Discussion
----------
[Process] Ecaping of CLI arguments containing slashes on Windows
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #22549
| License | MIT
Actually only the first argument - the command needs to be escaped but that would need another condition. I think it should be OK.
Commits
-------
0d073128de [Process] Ecaping of CLI arguments containing slashes on Windows
This PR was merged into the 3.3-dev branch.
Discussion
----------
Not allowing autoconfigure, instanceofConditionals or defaults for ChildDefinition
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | yes (removing risky behavior)
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | see #22530
| License | MIT
| Doc PR | n/a
This PR *prohibits* using `autoconfigure`, `_instanceof` and `_defaults` for ChildDefinition.
Additionally, I added many "integration" test cases: we need to test and prove all edge cases. These are in the `integration/` directory: the `main.yml` file is parsed and compared to `expected.yml`. Both are in YAML to ease comparing the before/after. We need to check these out and make sure they're right and we're not missing anything else.
This PR removes MANY of the "wtf" cases, but there are still 4 that I know of... and of course they all deal with parent-child stuff :).
A) [MAJOR] [autoconfigure_parent_child_tags](https://github.com/symfony/symfony/pull/22563/files#diff-fd6cf15470c5abd40156e4e7dc4e7f6d) `instanceof` tags from autoconfigure are NEVER applied to the child (you can't set `autoconfigure` directly on a Child, but you still can set it on a parent and inherit it... sneaky). We could throw an Exception I suppose to prevent this `autoconfigure` from cascading from parent to child... but it's tricky due to `instanceof`.
B( [MAJOR] [instanceof_parent_child](https://github.com/symfony/symfony/pull/22563/files#diff-14666e9a25322d44b3c2c583b6814dc2) `instanceof` tags that are applied to the parent, are not applied to the child. Again, you can't set `instanceof` directly on a Child, but you *can* set it on a parent, and have that cascade to the child. Like before, we could maybe throw an exception to prevent this.
C) [MINOR] ([autoconfigure_child_not_applied](https://github.com/symfony/symfony/pull/22563/files#diff-3372a1dcaf3af30d14a7d0a6c8bfa988)) automatic `instanceof` will not be applied to the child when the parent class has a different (non-instanceof-ed) class. If we could throw an exception for (A), then it would cover this too.
D) `_tags` from defaults are never used (unless you have inherit_tags) - fixed in #22530
A, B & C are effectively caused by there being a "sneaky" way to re-enable `autoconfigure` and `instanceof` for ChildDefinition... which opens up wtf cases.
## Wait, why not support `_defaults`, `autoconfigure` and `_instanceof` for child definitions?
1 big reason: reduction of wtf moments where we arbitrarily decide override logic. PLUS, since `_defaults`, `instanceof` and `autoconfigure` *are* applied to parent definitions, in practice (other than tags), this makes no difference: the configuration will still pass from parent down to child.
Also, using parent-child definitions is already an edge case, and this *simply* prevents *just* those services from using the new features.
## Longer reasons why
The reason behind this is that parent-child definitions are a different mechanism for "inheritance"
than `_instanceof` and `_defaults`... creating some edge cases when trying to figure out which settings "win". For example:
```yml
# file1.yml
services:
_defaults:
public: false
ChildService:
parent: parent_service
# file2.yml
services:
_defaults:
public: true
ParentService: ~
```
Is `ChildDefinition` `public: true` (so the parent
overrides the child, even though it only came from _defaults) or `public: false` (where
the child wins... even though it was only set from its _defaults)?
Or, if ParentService is explicitly set to `public: true`, should that override the `public: false` of ChildService (which it got from its `_defaults`)? On one hand, ParentService is being explicitly
set. On the other hand, ChildService is explicitly in a file settings `_defaults` `public: false`
There's no correct answer.
There are also problems with `_instanceof`. The importance goes:
> defaults < instanceof < service definition
But how do parent-child relationships fit into that? If a child has public: false
from an _instanceof, but the parent explicitly sets public: true, which wins? Should
we assume the parent definition wins because it's explicitly set? Or would the
_instanceof win, because that's being explicitly applied to the child definition's
class by an _instanceof that lives in the same file as that class (whereas the parent
definition may live in a different file).
Because of this, @nicolas-grekas and I (we also talked a bit to Fabien) decided that
the complexity was growing too much. The solution is to not allow any of these
new feature to be used by ChildDefinition objects. In other words, when you want some
sort of "inheritance" for your service, you should *either* giving your service a
parent *or* using defaults and instanceof. And instead of silently not applying
defaults and instanceof to child definitions, I think it's better to scream that it's
not supported.
Commits
-------
a943b96d42 Not allowing autoconfigure, instanceofConditionals or defaults for ChildDefinition
This PR was merged into the 3.3-dev branch.
Discussion
----------
[Console][HttpKernel] Avoid reflection-based registration for command public services
| Q | A
| ------------- | ---
| Branch? | 3.3
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | https://github.com/symfony/symfony/pull/22410#issuecomment-298158585
| License | MIT
| Doc PR | n/a
By mapping commands ids by their alias in `console.command.ids` (even if the alias is not registered in the container for public services), then skipping reflection if the predictable alias exists as a key of `console.command.ids`.
Please note that the whole command service registration process is far from ideal.
I'm working on changing this for 3.4 in a transparent way regarding end users, leveraging PSR-11 to make the console component DI friendly, allowing to register commands as true private services (no more public aliases) and providing laziness for those.
Commits
-------
6c1b384b75 Avoid reflection-based registration for command public services
This PR was squashed before being merged into the 3.3-dev branch (closes#22564).
Discussion
----------
Fixing problem where _defaults set to null was seen as a service
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
```yml
services:
_defaults:
```
If you leave `_defaults` empty (i.e. null), you got a bad error before. Now it's better :)
Before:
>The definition for "_defaults" has no class. If you intend to inject this service dynamicall
y at runtime, please mark it as synthetic=true. If this is an abstract definition solely use
d by child definitions, please add abstract=true, otherwise specify a class to get rid of th
is error.
After:
> Service "_defaults" key must be an array, "NULL" given in "/path/to/services.yml"
Commits
-------
4b7e148a9b Fixing problem where _defaults set to null was seen as a service
Also, not allowing arguments or method calls for autoconfigure. This is a safety
mechanism, since we don't have merging logic. It will allow us to add this in the
future if we want to.
The reason is that parent-child definitions are a different mechanism for "inheritance"
than instanceofConditionas and defaults... creating some edge cases when trying to
figure out which settings "win". For example:
Suppose a child and parent definitions are defined in different YAML files. The
child receives public: false from its _defaults, and the parent receives public: true
from its _defaults. Should the final child definition be public: true (so the parent
overrides the child, even though it only came from _defaults) or public: false (where
the child wins... even though it was only set from its _defaults). Or, if the parent
is explicitly set to public: true, should that override the public: false of the
child (which it got from its _defaults)? On one hand, the parent is being explicitly
set. On the other hand, the child is explicitly in a file settings _defaults public
to false. There's no correct answer.
There are also problems with instanceof. The importance goes:
defaults < instanceof < service definition
But how does parent-child relationships fit into that? If a child has public: false
from an _instanceof, but the parent explicitly sets public: true, which wins? Should
we assume the parent definition wins because it's explicitly set? Or would the
_instanceof win, because that's being explicitly applied to the child definition's
class by an _instanceof that lives in the same file as that class (whereas the parent
definition may live in a different file).
Because of this, @nicolas-grekas and I (we also talked a bit to Fabien) decided that
the complexity was growing too much. The solution is to not allow any of these
new feature to be used by ChildDefinition objects. In other words, when you want some
sort of "inheritance" for your service, you should *either* giving your service a
parent *or* using defaults and instanceof. And instead of silently not applying
defaults and instanceof to child definitions, I think it's better to scream that it's
not supported.
This PR was squashed before being merged into the 3.3-dev branch (closes#22531).
Discussion
----------
Throwing an exception if the class is not found
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
We now throw an exception if the user makes a mistake with their PSR-4 prefix and namespace. For example:
```yml
AppBundle\Controller\:
resource: '../../src/AppBundle/{Controller}'
public: true
```
I should not have the `\Controller` at the end of the key. Previously, it would silently not import any services from the directory. Now it throws:
> Expected to find class "AppBundle\Controller\Controller\Admin\BlogController" in file "/path/to/project/src/AppBundle/Controller/Admin/BlogController.php" while importing services from resource "../../src/AppBundle/{Controller}", but it was not found! Check the namespace prefix used with the resource.
The only "downside" is that this prevents someone from importing files from a resource that has a file with no class in it (`functions.php`). @nicolas-grekas and I decided today that we can throw an exception now to be safe, and see if anyone has that valid use-case.
Cheers!
Commits
-------
e85bcc9 Throwing an exception if the class is not found
This PR was merged into the 3.3-dev branch.
Discussion
----------
[FrameworkBundle] Restore 3.2-like behavior for FormPass, to help BC with Sonata
| Q | A
| ------------- | ---
| Branch? | 3.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
I tried updating a Sonata project to 3.3, and found it broken.
The issue is that Sonata uses the constructor arguments of the `form.extension` to create its own `form.extension` service - but borrows its first args from the Symfony one.
Here is the form pass doing that:
https://github.com/sonata-project/SonataCoreBundle/blob/3.x/DependencyInjection/Compiler/FormFactoryCompilerPass.php
And the implementation of the form extension:
https://github.com/sonata-project/SonataCoreBundle/blob/3.x/DependencyInjection/SonataCoreExtension.php
Question: is this covered by the BC policy? It shouldn't to me, because that would prevent *any* service reconfiguration.
Thus, I'm proposing the attached patch, which basically reverts the deprecated `FormPass` in FrameworkBundle to its 3.2 state.
I added a check to the new `FormPass` in the Form component so that it doesn't overwrite such compatibility configurations.
See for corresponding fix on https://github.com/sonata-project/SonataCoreBundle/pull/399
Commits
-------
c97b08e6c0 [FrameworkBundle] Restore 3.2-like behavior for FormPass, to help BC with Sonata
This PR was merged into the 3.3-dev branch.
Discussion
----------
[SecurityBundle] Enhance FirewallContext::getListeners()
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | https://github.com/symfony/symfony/pull/20417#discussion_r91704023, https://github.com/symfony/symfony/pull/20417#discussion_r91704145
| License | MIT
| Doc PR | symfony/symfony-docs#... <!--highly recommended for new features-->
I think @stof is right.. and the fact we can do this on master currently without the hassle.
cc @chalasr
Commits
-------
ba650783f5 [SecurityBundle] Enhance FirewallContext::getListeners()
This PR was merged into the 2.7 branch.
Discussion
----------
[Asset] Preventing the base path or absolute URL from being prefixed incorrectly
| Q | A
| ------------- | ---
| Branch? | 2.7
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
Fixes an edge case (which I need) where the version strategy returns an absolute URL. Currently, if this happens, the baseUrl or basePath is prefixed - giving `https://baseurl.com/https://pathreturnedfromversioning.com` or `/basePath/https://pathreturnedfromversioning.com`.
I don't see any reason to prevent an absolute URL from being returned by the version strategy. And it's not a BC break, because the previous paths that were returned were nonsense.
Cheers!
Commits
-------
746c91eea4 Preventing the base path or absolute URL from being prefixed incorrectly on an absolute URL
* 3.2:
Fixed the flickering when loading complex profiler panels
[Console] Fix bar width with multilines ProgressBar's format
[DI] Add missing check in PhpDumper
[Serializer] XmlEncoder: fix negative int and large numbers handling
[Console] Fix dispatching throwables from ConsoleEvents::COMMAND
* 2.8:
Fixed the flickering when loading complex profiler panels
[DI] Add missing check in PhpDumper
[Serializer] XmlEncoder: fix negative int and large numbers handling
[Console] Fix dispatching throwables from ConsoleEvents::COMMAND
* 2.7:
[DI] Add missing check in PhpDumper
[Serializer] XmlEncoder: fix negative int and large numbers handling
[Console] Fix dispatching throwables from ConsoleEvents::COMMAND
This PR was merged into the 3.3-dev branch.
Discussion
----------
[Console] Review console.ERROR related behavior
| Q | A
| ------------- | ---
| Branch? | 3.3
| Bug fix? | yes
| New feature? | yes
| BC breaks? | no
| Deprecations? | yes
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
This PR is a follow up of #18140 that I wanted to do since a few weeks.
It enhances this work with fixes and behavior changes.
It embeds #22435 and resolves issues like the one described in #20808.
- makes ConsoleErrorEvent *not* extend the deprecated ConsoleExceptionEvent
- replace ConsoleErrorEvent::markErrorAsHandled by ConsoleErrorEvent::setExitCode
- triggers the deprecation in a more appropriate moment
- renames ExceptionListener to ErrorListener
- tweaks the behavior in relation to #22435
Commits
-------
a7c67c9ab2 [Console] Review console.ERROR related behavior
This PR was merged into the 3.3-dev branch.
Discussion
----------
[Cache] Keep only hit/miss (not values) in TraceableAdapter/Cache
| Q | A
| ------------- | ---
| Branch? | 3.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Right now, TraceableAdapter and TraceableCache both keep all fetched values.
This exhaustive reporting is too much data gathering to me.
Here is a PR that keeps only true/false for each hit/miss + all corresponding keys.
That can still be a lot of data (one item per fetched key) - but a bit better.
Should we go further in stats-only gathering? Or should we *not* do this and keep collecting *all* the values?
The PR also fixes "Traversable" handling.
ping @Nyholm
The profiler panel still works (although breaking lines in the middle of words is strange, but this is the profiler's CSS, nothing special to this specific case).
Commits
-------
0c73c5d [Cache] Keep only hit/miss (not values) in TraceableAdapter/Cache
This PR was squashed before being merged into the 3.2 branch (closes#21958).
Discussion
----------
[Console] Fix bar width with multilines ProgressBar's format
| Q | A
| ------------- | ---
| Branch? | 3.2
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
The bar width is badly recalculated when we use multilines (``\n``) and bar placeholer (``%bar%``) in ProgressBar's format. The bar width is reduced because multilines is considered as a big oneline. This PR fixes this.
Commits
-------
3d58ab5bad [Console] Fix bar width with multilines ProgressBar's format
